The twenty-five-year time period from 2000 to 2025 was for sure one of the most remarkable economic changes that has occurred in modern history, with the country of China changing drastically from an inferior lower-middle-income country to a much more upper-middle-income nation that is close to the threshold of high-income status. This journey has been filled with so many key but fundamental changes in terms of economic structure, growth drivers, and policy ideas that are calling for a lot more systematic scholarly research work and even investigation. As China goes into the twenty-first century, its economy is mainly being affected and shaped by investment-driven expansion growth and even lots much more uses in so many global markets after the World Trade Organization rises back in the year of 2001. By 2025, however, the economic scenes had changed quite very dramatically, with services contributing more than half of gross domestic product, technology-intensive manufacturing gaining so much more status and very high prominence, along with other kinds of policy frameworks that are strongly focusing and stressing on having a lot more high-quality development instead of just some kind of normal quantitative expansion. Understanding this kind of change is not only some sort of easy exercise and work in terms of historical documentation but also has so many more profound implications and even effects for other development economics out in the real world, as China’s experience gives us both useful and highly theoretical insights, ideas and practical lessons for other newly rising economies also dealing with structural transformation. The existing literature has very extensively recorded of and discussed about China’s growth miracle, yet there are so many more gaps remaining in terms of understanding the so many interactions between industrial evolution, productivity dynamics, and policy interventions in many more extended temporal areas. This study deals with these sorts of gaps by systematically analyzing the changes involved in China’s economic changes from 2000 to 2025, with lots of attention given out to the changing contributions of primary, secondary, and tertiary areas and sectors to growth, before making more evidence-based projections for the 2026-2035 period. The research is being guided by three connected goals and objectives: first, to quantify and describe the structural changes of China’s economy through very detailed analysis of area level contribution patterns; second, to go over the drivers of productivity evolution in different manufacturing and service sectors, with special attention to the issue of service-manufacturing integration; and third, to forecast on better and more plausible growth ways there are in the real world and industrial development pathways for the new coming decade in term of rising and emerging technologies and policy goals. By combining this very rigorous quantitative analysis with today’s theoretical frameworks and models from development economics and industrial organization, this study will contribute to scholarly understanding of how large developing economies deal with the transitions and changes from factor-driven to other kinds of innovation-driven growth models.

The study proceeds by structuring empirical documentations of China’s structural changes covering the 2000-2025 period through shift-share analysis and Levinsohn-Petrin TFP ([8]) estimations, culminating in ARIMA-based scenario projections for the 2026-2035 period. The empirical results demonstrate that while investment-led growth generated crowding-out effects, service-manufacturing integration offers a viable pathway to resolve [2] disease and achieve future growth targets through digital-green synergies.

First of all, the theoretical foundations and basis for understanding China’s economic change rely on several complementary but related intellectual traditions within our development economics and growth theory. The structural change of literature, first coming off from the pioneering work of [6] and [5], gives us very essential, key and useful conceptual tools for analyzing the actual redistribution and even reallocation of all these kinds of economic activity in all these different areas and fields during our development. These frameworks are basically arguing that sustained growth often involves the weakening and declining share of agriculture in terms of issues like output and employment, the hump-shaped ways of manufacturing, and finally the later rises of other kinds of services as economies reach more and more closer to their own levels of maturity. [9] dual-sector model, with its focus on labor transfer from traditional to modern sectors, continues to inform on these so called understandings of China’s rural-urban migration and shifts and industrialization processes, though lots of other scholarship has highlighted and shown about the model’s limitations in term of getting back and even capturing institutional issues out there that are being specific to China’s context and background. More recently, [1] famous work on institutional foundations of prosperity has also brought out lots of arguments and even debate about the very role of political-economic arrangements when it comes to affecting and shaping China’s very own developmental ways. The changes of growth theory from exogenous to endogenous frameworks have also affected and even shaped scholarly methods and even approaches to China’s transformation, with [12] focus on technological change as an endogenous outcome or result in this case of profit-motivated innovation finding particular connection in analyses of China’s transition from imitation to innovation strategies.

Meanwhile, empirical research on China’s area level changes and even transformation has produced quite rich but sometimes contradictory findings that call for much more careful synthesis and work to be done. Studies going over the pre-2000 period, like those by [10] and [7], discussed about the initial phases of market liberalization and their differential impacts in other sectors. The post-WTO accession era got back very extensive scholarly attention, with [3] editing a comprehensive volume that goes over China’s great economic changes through many analytical lenses. Research specifically addressing sectoral contributions to growth has benefited from improved and bettered data availability and methodological work. [4] made use of input-output analysis to track down the changing interconnections between many manufacturing and services, finding increasing interdependence that made simple sectoral classifications more involved here. [13] NBER analysis gave us very useful and valuable insights and knowledge about the investment-driven nature of China’s growth between 1998 and 2015, and recorded that capital deepening contributed and even led to 68.3 percent of GDP growth for each worker during this time period, while total factor productivity’s contribution declined and went down from its main and key position in earlier times and decades. This finding bonds very nicely with other larger issues and concerns about the actual level of sustainability of investment-led growth models and the need of going toward productivity-driven expansion.

Moreover, service sector development has gotten back more and more increasing scholarly attention as its economic significance has grown more and more. The service sector’s share of GDP surpassed manufacturing in 2012, marking a huge structural milestone that positioned China in our global patterns of tertiarization. However, this expansion has been accompanied by other concerns about productivity performance, with the idea of “Baumol’s disease” featuring very in front in many scholarly debates. [14] had already given us a lot more high-quality and very rigorous empirical analysis work about this very kind of issue existing in China, and were arguing that service sector productivity has definitely done negatively and even lagged behind manufacturing, creating what they call as a divergent sort of industrial transformation with rising status of services alongside relatively low service sector productivity. Their research, making use of China’s tax survey data and OECD input-output tables, shows that increasing service sector input to manufacturing can improve the relative productivity of services and help overcome the Baumol’s disease trap, with producer services playing a very significant role in this process. This finding argues that service-manufacturing integration was a key pathway for solving productivity issues found in structural transformations.

This study uses a mixed-methods research design that combines quantitative analysis of historical data with model-based forecasting, justified by the need to both record of past patterns and project future trajectories while triangulating findings in many or multiple approaches. The quantitative component made use of time-series data from the National Bureau of Statistics of China covering 2000 to 2025, supplemented by data from the OECD Input-Output Database, World Development Indicators, and China Household Finance Survey. See Figure 1.

Figure 1.Time-series datasets and numerical values from the National Bureau of Statistics of China covering the actual years of 2000 to 2025.

The National Bureau of Statistics gives us the primary source for GDP series by expenditure and production approaches, sectoral value-added, and other related employment data, with consistency checks being done as well to get back more comparability in different methodological revisions. See Figure 2.

Figure 2.OECD-ICIO tables from the National Bureau of Statistics of China.

The OECD-ICIO tables allow for a comprehensive analysis of cross-sectoral connections, particularly the service-manufacturing integration index (defined as the proportion of service sector inputs embedded in total manufacturing output). The China Household Finance Survey data, as it was being collected by Southwestern University of Finance and Economics from before, also gave us quite useful micro-level information on issues like consumption, housing, and wealth that show off key information about distributional aspects of structural changes we have here.

To ensure the empirical base is transparent, the analysis relies on four specific datasets: 1) Enterprise tax survey data, utilizing firm-level units of analysis with a sample of manufacturing and service enterprises from 2000-2025 to track micro-level productivity; 2) A cross-country panel dataset constructed from OECD and World Bank sources, covering country-level metrics for 35 emerging and advanced economies over 2000-2025; 3) The China Household Finance Survey, employing household-level microdata (approx. 40,000 households per wave) to assess the distributional effects of housing policies; and 4) A regional pilot sample covering ten designated municipal zones to evaluate the policy effects of digital-green initiatives.

The analytical framework consisted of three interconnected components designed to address the research objectives systematically. The core model specifications are strictly defined: 1) In the Levinsohn-Petrin TFP estimation, the dependent variable is firm-level value-added, with capital and labor as key regressors, and intermediate inputs used to control for unobservable productivity shocks. 2) For the cross-country panel analysis, service productivity is the dependent variable, the service-manufacturing integration index is the key regressor, with trade openness and human capital as controls, utilizing country and year fixed effects. 3) The ARIMA model utilizes historical annual GDP growth rates as the primary time-series input to forecast future trajectories. See Figure 3.

First, descriptive statistical analysis tracks down the evolution of sectoral shares in terms of GDP and employment, calculating contribution coefficients (defined as the ratio of a specific sector’s value-added increment to the total GDP increment over a given period) to measure each sector’s proportional contribution to overall growth. See Figure 4.

Sectoral value-added and employment data are analyzed using shift-share decomposition ([11]) to isolate within-sector productivity improvements from labor reallocation effects. Concurrently, drawing on the approach of [14], this study adopts various types of enterprise-level tax survey data to conduct robust estimations of total factor productivity using the Levinsohn-Petrin methodology, which effectively addresses various biases in production function estimation.

Figure 5 exhibits cross-country panel analysis to compare Chinese economy in global perspective. It can be seen that the increase and development of Chinese economy take lead globally.

Figure 3.Research analytical framework.

Figure 4.Descriptive statistical analysis charts.

Figure 5.Cross-country panel analysis charts.

Figure 6.ARIMA architecture model.

To project future economic trends, this study establishes an Autoregressive Integrated Moving Average (ARIMA) framework. The forecasting design includes a short validation step, comparing ARIMA-fitted values against actual 2015-2025 historical data to confirm predictive accuracy. Projections are structured around three scenarios: a baseline scenario assuming current policy trajectories and moderate technological adoption; an upside scenario assuming accelerated digital-green synergies and deep structural reforms; and a downside scenario modeling delayed transitions and persistent property sector drag. The critical threshold of a 5.77% average annual growth requirement is mathematically derived from the national strategic target to double the 2020 per capita GDP by the year 2035, serving as the benchmark for evaluating these scenarios. See Figure 6. The ARIMA method had gotten back so many time-series properties and attributes of GDP and sectoral growth rates and this is also allowing for the very uses of outside information about things like policy targets and technological trends in our world.

To contextualize the overarching economic trajectory, China’s GDP growth exhibited a clear deceleration from an annual average of 10% in the 2000s to approximately 5% by the early 2020s, as shown in Figure 7. This initial rapid expansion was predominantly investment-driven, with capital deepening accounting for 68.3% of per-worker growth between 1998 and 2015, significantly overshadowing total factor productivity gains, as shown in Figure 8. While this investment-led model facilitated rapid short-term scaling, it also generated structural vulnerabilities, particularly within the real estate sector. Boom-and-bust cycles in housing were closely tied to shadow banking mechanisms, where tightening monetary policies inadvertently stimulated alternative financing channels and overheated the market.

Figure 7.China’s GDP growth from 2000 to 2025.

Figure 8.Investment-driven nature of growth during 1998-2015.

Concurrently, the economic structure underwent significant realignments. The primary sector’s GDP share declined to 7% by 2025, and the secondary sector contracted to 38%, albeit with internal upgrading toward high-tech manufacturing, as shown in Figure 9. The tertiary sector emerged as the dominant economic driver, surpassing 55% of GDP, as illustrated in Figure 10. However, the disproportionate expansion of lower-value consumer services relative to high-tech business services poses the risk of Baumol’s disease, as service productivity has historically lagged

Figure 9.Sectoral contribution analysis.

Figure 10.Tertiary sector’s ascendancy.

behind manufacturing. Overcoming this productivity stagnation requires deepening service-manufacturing integration and leveraging emerging digital and green technology synergies, as shown in Figure 11, a pathway that international comparisons suggest is crucial for avoiding premature deindustrialization and sustaining long-term economic health, as shown in Figure 12.

Figure 11.Changing role of digital and green sectors.

Figure 12.International comparison.

Shift-share decomposition reveals that total labor productivity growth decelerated from an annual average of 8.2% (2000-2010) to 6.4% (2010-2020), and ultimately to 4.1% (2020-2025). Sectoral contribution patterns indicate a profound structural shift: the primary sector’s contribution to growth fell from 4.2% to 2.5%, the secondary sector declined from 52.3% to 35.8%, and the tertiary sector expanded dramatically from 43.5% to 61.7%. Within manufacturing, high-tech and equipment manufacturing expanded to 17.1% and 36.8%, respectively, whereas traditional manufacturing contracted. Concurrently, service sector productivity grew at merely 1.2% annually compared to 3.8% in manufacturing, expanding the productivity gap and highlighting risks of Baumol’s disease. However, instrumental variable analysis yields a coefficient of 0.324, demonstrating that increased service inputs into manufacturing significantly mitigate this productivity lag. To address endogeneity, historical input-output coefficients from the year 2000 are used as the instrument. This satisfies the relevance condition (historical I-O patterns strongly predict current integration) and the exclusion restriction (historical structures affect current productivity only through the persistent integration channel).

A difference-in-differences (DID) analysis evaluates the ten regional pilot programs. Treated regions were selected based on exogenous national policy designations rather than pre-existing economic performance, and event-study checks confirmed parallel pre-trends between treated and control regions prior to policy implementation. Post-treatment, treated regions had an 8.2 percentage point higher digital adoption rate (p < 0.01) and a 5.7 percentage point higher renewable energy share (p < 0.05) compared to control regions. The synergy effect—as it’s being measured as very interaction between digital and green investments—is being very positive and very significant: this was a 10% increase in terms of digital investment along with 10% increase in green investment that gives us back a 3.2% extra productivity gain beyond just total and the sum and total of individual effects (p < 0.01). Hebei’s focus on digital infrastructure coordinated with renewables has basically gotten back a PUE of 1.38 (vs national average 1.44), while Shenzhen’s digital empowerment approach has had higher digital adoption rates of 82% (vs national average 68%).

The ARIMA model selection procedure was being done using AIC and BIC criteria, which selects an ARIMA(1, 1, 2) specification for GDP growth, with autoregressive coefficient 0.74 (s.e. 0.08) and moving average coefficients −0.32 and −0.18 (both significant at p < 0.05) in this case right over here.

The empirical findings confirm that China’s 1998-2015 investment-led model, where capital deepening drove 68.3% of per-worker growth compared to 31.7% from TFP, generated structural imbalances. Notably, a 10% increase in infrastructure stimulus led to a 4.2% credit crowding-out effect on private enterprises. Addressing Baumol’s disease requires crossing a critical integration threshold; elevating China’s service-manufacturing integration index from 0.42 to the 0.45 threshold observed in advanced economies could boost service productivity by 15%. Furthermore, housing market interventions—such as the 2014 mortgage easing—yielded negative long-term consumption effects (totaling −2.1% over five years) and exacerbated generational wealth disparities.

There are also some other kinds of limitations that are qualifying our findings. First, the Levinsohn-Petrin TFP estimation, while addressing on issue of simultaneity, cannot fully control for other new price differences in sectors that may bias productivity comparisons. Service sector output measurement remains as being very problematic, particularly in digital platforms where user value making is not collected in revenues. Second, our housing market analysis relies too much on transaction data that may overrepresent formal market activity, missing the key and main shadow banking and informal financing channels. Third, the regional pilot analysis suffers quite badly from potential selection bias. Regions chosen for pilots may have had pre-existing advantages that confound treatment effect estimates.

These limitations over here also suggest productive avenues for future research. First, making more and more better service sector productivity measures that account for quality change and digital platform economics would improve our understanding of Baumol’s disease changes here. Second, micro-level analysis of firm-level service-manufacturing use—tracking down individual firms’ diversification into services and service firms’ specialization in producer services—would show the mechanisms behind these aggregate panel results. Third, housing market research should have shadow banking flows into general equilibrium models to get the full range of financial sector interactions. Fourth, digital-green research needs firm-level data on technology adoption and energy efficiency investments to find other causal mechanisms. Fifth, forecasting would benefit from machine learning approaches that can get us so many nonlinearities and threshold effects that linear ARIMA models miss out on.

This research has conducted a systematic empirical analysis of China’s economic changes from 2000 to 2025 and utilized forecasting models to project growth trajectories for the 2026-2035 period. The main findings show a quite clear pattern. Farming became a much smaller part of the economy, manufacturing grew first, then shrank, and services ended up being the biggest piece of the pie. In this big picture, a few things really stand out. From 1998 to 2015, growth was driven mostly by investment, which worked very fast but also created some long-term and more lasting problems. Service companies never quite caught up with factories in terms of issues like productivity, which raises the risk of what economists call Baumol’s disease. Housing went through big boom and bust cycles that hit so many different groups of people in very different ways. At the same time, digital technology and green energy became key, main and major policy priorities for the future. The numbers show that when factories and services work together more and more closely, it helps solve many productivity problems. They also show that changes in housing prices have long-lasting effects on what people can spend and how unequal things become. Reaching the goals set for 2035 will take steady gains in terms of work productivity, and that will need both new technology and real changes in how things are being set up.

Of course, no study is perfect here, and this one has its limits as well. Using official numbers from the government is needed, but those numbers can have mistakes or biases, especially when going back to earlier years. The company-level data we used comes from tax records, which are rich and useful, but they miss all the informal business activity that still matters in some service areas. Looking ahead is always tricky. Even though we tried different case scenarios and checked our work carefully, there is just a lot we cannot know about areas like future technology, policy choices, and world conditions. Going forward, researchers should dig deeper into how services and factories connect at smaller scales, both by industry and by region. They should also look closely at exactly how digital tools change productivity in different kinds of companies. And they need to keep studying who wins and who loses from these big economic shifts, in different places, age groups, and income levels. There is still a lot more to figure out.